As with every release of Microsoft Dynamics GP, our test team has been hard at work creating the latest and greatest performance benchmark tests to help illustrate the incredible scalability and performance that Microsoft Dynamics GP can deliver. The latest benchmarks prove yet again that even under the most strenuous of environments, Microsoft Dynamics GP continues to perform!
Recently, the Microsoft Dynamics GP team completed 3 separate performance benchmark tests. They were:
1. 500 constant concurrent users in the rich Microsoft Dynamics GP client, all engaged in heavy transaction processing or analysis activity across various functional areas of the solution.
2. 500 constant concurrent users in the rich Microsoft Dynamics GP client in addition to 10 instances of continuous (synchronous) integrating Web Services clients creating Sales Order transactions.
3. US Payroll testing, including both check processing and payroll reporting for a 30,000 employee organization.
As you can see from the results below, the performance benchmark numbers are very impressive...
Transactions per 8 hour day
Transaction Lines per 8 hour day
1. 500 concurrent users
2.96 million transactions
9.85 million transaction lines
2. 500 concurrent users + 10 Web Service instances
3.20 million transactions
11.29 million transaction lines
Time to complete
3. 30,000 employee check run & reporting
· Build, Calculate, Print, and Post 30,000 checks: 8 Hours
· Month-End Reporting: 1.56 Hours
· Year-End Reporting: 49.4 Minutes
· Total time for check processing and reporting: 10.4 Hours
Customers and Partners can download the complete White Paper detailing these three tests and their results here: https://mbs.microsoft.com/customersource/documentation/whitepapers/MSD_GP2010ScalabilityPerfomance
500 vs. 1,000 Concurrent Users
Those with knowledge of Microsoft Dynamics GP's history around benchmark testing may question why this test focused on 500 concurrent users as opposed to the previous 1,000 concurrent user tests. The answer is quite simple. Previous benchmark tests for Microsoft Dynamics GP were done in a 1,000 concurrent user environment, and were focused on proving the user scalability potential of the system. Based on real-world customer feedback, Microsoft’s goal has shifted to focus more on transaction throughput in a user environment more in-line with Microsoft Dynamics GP’s target market. In addition to this change in focus, a new test has been added to simulate common real-world environments where customers are leveraging Web Services for Microsoft Dynamics GP to import data from external systems while users engage in heavy transaction processing. As a result of this change in focus and the continuing evolution of the testing hardware and environment, readers should not compare outcomes of these tests against previous versions of performance benchmark results.
We hope you enjoy this latest testament to the great scalability of Microsoft Dynamics GP!
I'm not encouraged by the payroll results. I don't think 8 hours is acceptable here. If you have to verify results or, heaven forbid recalculate or rebuild this will put stress on the payroll team and push them into overtime or the next day. I have a client that pays 1000 employees and their process goes into the next day. In my opinion GP's payroll processing performance requires enhancement.
Thanks for the feedback Mike. There are a few things I would like to mention to help here.
The time required to run payroll is highly dependent on two primary things: Hardware, and complexity of your payroll data (number of pay records, deductions, benefits, etc.). If you have a client that is taking longer than a day to pay 1,000 employees, it sounds like there could be issues either with the hardware or with the setup/processes in the payroll system. We have many customers that are paying 1,000+ employees and the feedback we here is this takes less then a morning to complete the process.
With GP 2010, you can have multiple users in payroll building, calculating, and processing check batches, so this can definitely cut down on issues when you run into an error on a large batch and end up having to rebuild and calculate it again.
It may be worthwhile to consider engaging our support team to help your client diagnose if there are any issues that are slowing down their payroll processes. They are very good at isolating hardware and/or setup issues that impact performance.
Thanks again for the feedback!